# Anyone tried Zola with a huge site?

**URL:** <https://zola.discourse.group/t/anyone-tried-zola-with-a-huge-site/951>\
**Category:** Support\
**Created:** [June 20, 2021, 8:56pm UTC](https://zola.discourse.group/t/anyone-tried-zola-with-a-huge-site/951 "2021-06-20T20:56:33Z")\
**Posts on this page:** 6\
**Page:** 1

<div class="post-metadata">

**Author:** ![probable-potato](https://yyz2.discourse-cdn.com/free1/user_avatar/zola.discourse.group/probable-potato/32/594_2.png) [@probable-potato](https://zola.discourse.group/u/probable-potato)\
**Post date:** [June 20, 2021, 8:56pm UTC](https://zola.discourse.group/t/anyone-tried-zola-with-a-huge-site/951/1 "2021-06-20T20:56:33Z")

</div>

I’m contemplating using Zola for a personal intranet site that could potentially have tens of thousands of pages. (Let’s say, thinking ahead, 100K pages.) Has anyone tried something like this? Am I crazy?

---

<div class="post-metadata">

**Author:** ![keats](https://avatars.discourse-cdn.com/v4/letter/k/34f0e0/32.png) [@keats](https://zola.discourse.group/u/keats)\
**Post date:** [June 21, 2021, 8:59am UTC](https://zola.discourse.group/t/anyone-tried-zola-with-a-huge-site/951/2 "2021-06-21T08:59:17Z")

</div>

You can try it: [zola/gen.py at master · getzola/zola · GitHub](https://github.com/getzola/zola/blob/master/components/site/benches/gen.py#L172)

It should be ok but there are currently some performance issues with taxonomies I think which I haven’t had the time to look into so if you’re using them heavily it could be slow.

---

<div class="post-metadata">

**Author:** ![mwcz](https://yyz2.discourse-cdn.com/free1/user_avatar/zola.discourse.group/mwcz/32/579_2.png) [@mwcz](https://zola.discourse.group/u/mwcz)\
**Post date:** [May 3, 2022, 4:35pm UTC](https://zola.discourse.group/t/anyone-tried-zola-with-a-huge-site/951/3 "2022-05-03T16:35:30Z")

</div>

I’m pretty new here but I’m using Zola to build a site with about 140,000 pages currently. I’m not using taxnonomies, but I am using syntax highlighting, and the site builds in about 20 seconds with a pretty fast SSD and a 5900x. I also evaluated Hugo and several JS based engines and Zola was the fastest. It was about 4x faster than Hugo and 100x or more faster than the JS engines.

---

<div class="post-metadata">

**Author:** ![keats](https://avatars.discourse-cdn.com/v4/letter/k/34f0e0/32.png) [@keats](https://zola.discourse.group/u/keats)\
**Post date:** [May 3, 2022, 4:46pm UTC](https://zola.discourse.group/t/anyone-tried-zola-with-a-huge-site/951/4 "2022-05-03T16:46:13Z")

</div>

I didn’t expect the x4 with Hugo :o  
I’m wondering where are the slow areas, I’ve tried setting up some flamegraph but it doesn’t play well with the latest macos, I’ll need to create a linux VM again.  
Can you build your site with and without syntax highlighting to see the difference? Last time I did, it was almost 50% of the runtime. As well as the build time with current release version and current `next` branch.

---

<div class="post-metadata">

**Author:** ![mwcz](https://yyz2.discourse-cdn.com/free1/user_avatar/zola.discourse.group/mwcz/32/579_2.png) [@mwcz](https://zola.discourse.group/u/mwcz)\
**Post date:** [May 3, 2022, 9:14pm UTC](https://zola.discourse.group/t/anyone-tried-zola-with-a-huge-site/951/5 "2022-05-03T21:14:03Z")

</div>

Sure, I’ve been meaning to do that for a while now. I’m not sure when I’ll have the time but when I do, I’ll let you know what the results are!

---

<div class="post-metadata">

**Author:** ![mwcz](https://yyz2.discourse-cdn.com/free1/user_avatar/zola.discourse.group/mwcz/32/579_2.png) [@mwcz](https://zola.discourse.group/u/mwcz)\
**Post date:** [March 3, 2023, 2:13pm UTC](https://zola.discourse.group/t/anyone-tried-zola-with-a-huge-site/951/6 "2023-03-03T14:13:47Z")

</div>

Hey @keats, sorry I never did those measurements. Here’s some measurements with a one-week old build from `next` (`e42283e9`). Not much difference between syntax highlighting on & off.

139,781 pages

With syntax highlighting: 34,190 ms  
Without: 33,410 ms

There are hundreds of thousands of code blocks, however very few of them are tagged with syntax groups, which might explain why the timings are so close.
